A technique of detecting the presence of a line of cars which stop before a road with poor visibility such as a curve, a rising slope, and the like, and of reporting the detection result to a succeeding vehicle is useful from the viewpoint of preventing rear-end accidents. As one such technique, a technique of traveling vehicles tracking and detecting by using a milliwave radar arranged on a road is known. In this technique, first, a position and a speed of a vehicle are predicted by using tracking data obtained by a first scanning. Then, segmentation processing and attribute forming processing are performed on the data obtained by the first scanning and to the position-speed data obtained by the second scanning, and the result is collated, respectively. Further explanation is given for the technique of traveling vehicles tracking and detecting.
In the technique of traveling vehicles tracking and detecting, first, background subtraction processing is applied to measurement data sent from a milliwave sensor. This processing is the processing of obtaining the difference between the measurement data and the background data, deleting background fixed structures including power poles, guardrails, and the like, and leaving the measurement data of the moving objects. By making the background out of the target of the later mentioned processing by performing the background subtraction processing, more time for processing may be allocated for the processing of a moving object tracking and detecting.
Concerning generation of background data, a technique is known, for example, in which the background data are generated by updating the background data by taking a weighted average of the previous background data for measurement data. A technique is also known in which updating the background data is controlled in accordance with a detection result of whether or not a background object which has been temporarily obscured by other object is present.
The above-mentioned technique of traveling vehicles tracking and detecting, following the background subtraction processing, performs segmentation processing which groups moving objects with substantially the same speed and attribute forming processing which obtains attributes including a barycentric position, a speed, and the like of the groups from the measurement data of these groups. Then, on the basis of the obtained attributes, prediction of measurement data which are obtained next for the groups is performed, and by collating this prediction result and the actually obtained measurement data, tracking these groups is performed.
Techniques described in each of the following documents are known.
Document 1: Japanese Laid-open Patent Publication No. 2002-99986
Document 2: Japanese Laid-open Patent Publication No. 2000-172980
Document 3: Japanese Laid-open Patent Publication No. 2001-4742
In the detection of a moving object by a radar apparatus, a position of the moving object, that is, a distance from the radar apparatus, and a speed of the moving object, that is, a relative speed of the moving object with respecting to the radar apparatus, are obtained as measurement data. Here, when a rotary scanning type is used as the radar apparatus, a presence direction of the moving object seen from the radar apparatus is also obtained as measurement data.
FIG. 1 schematically illustrates a detection system which includes a rotary scanning type radar apparatus 1 and a controller 2, the detection system detecting a vehicle 4 which travels on a road 3. The radar apparatus 1 detects the presence direction of the vehicle 4 and a tree which is a structure 6 positioned on a road side 3 by scanning a detection area 5 rightwards and leftwards on a horizontal surface.
As mentioned above, when a rotary scanning type radar apparatus is used, the presence direction of the moving object is also obtained as measurement data, and therefore, the position of the moving object is specified in two dimensions. The rotary scanning type radar apparatus, however, is provided with a mechanical mechanism for rotating antennas in order to make a radar wave scan on the horizontal surface, and for such a mechanism, a maintenance operation of the radar apparatus is performed.
On the other hand, in a non-scanning type radar apparatus, the position of the moving object is specified in one dimension. Accordingly, when there is a plurality of objects present at the same distance from the radar apparatus, with the non-scanning type radar apparatus, the measurement data of these objects are not obtained for each object.
FIG. 2 schematically illustrates a detection system which includes a non-scanning type radar apparatus 1 and a controller 2, the detection system detecting a vehicle 4 which travels on a road 3. In such a radar apparatus 1 that does not scan the detection area 5, the detection of the presence direction of the vehicle 4 is not performed.
Here, a case is considered in which the above-mentioned technique of the traveling vehicle tracking and detecting is performed by using the non-scanning type radar apparatus. In this case, when the vehicle stops for a while at the same position as the structure present at a road side with respect to the distance from the radar apparatus for example, since the data of the vehicle have been lost by the above-mentioned background processing, it is difficult to perform the vehicle tracking and detecting.